Overview

While traditional keyword-based searches and conversational AI tools are effective for finding quick answers or general information, they fall short of facilitating deeper learning or providing structured, comprehensive insights. This project aims to revolutionize how people learn and grasp knowledge by introducing a system beyond traditional search engines and chatbot-based AI tools.

Project Details

  • Role

    UIUX Designer and Researcher

    Aug 2024 - Dec 2024

  • Team

    Eric Ye

    Emily Jeong

    Lillian Ding

    Developers: NOODEE Team

  • Tools

    Figma

    Google Suite

    Adobe Illustrator

    FigJam

Problem Statement

This was an ambitious task given our timeline of a few months, however, we were lucky enough to also know a few developers who wanted to see a change in GENAI development. Therefore, to bridge the gap of our vision to real work we first made a problem statement:

How might we design an AI-powered search tool that not only provides structured, deep knowledge but also enables users to explore, organize, and transform that knowledge into meaningful, practical outputs?

EARLY STAGES

Objective

The goal would be to create an experience that transforms the search engine into an effective educational tool, helping users gain deeper insights and systematically build knowledge, rather than just retrieving facts.

Target population

We first brainstormed a possible user group to focus on and from other platforms such as Milanote we took note of the users and decided to focus on:

Professionals, Students, and Content Creators

Study procedure

After brainstorming this user group we then thought about how we were going to obtain preliminary information even before designing. Because we want to improve current-day GENAI platforms we decided to examine user learning experiences through semi-structured interviews.

Moreover, because there are so many platforms that focus on GENAI to better learn the current issues and structure of the platforms we decided to also conduct a concept analysis of research based on these platforms.

Concept sketches

After examining various research papers on GENAI platforms and conducting some observational studies we created concept sketches of three concepts.

Concept 1:

Transforming browsing into a kind of digital treasure hunt. As users search for quick information, they can stumble upon hidden gems—interesting facts, historical details, or unexpected connections—making their exploration more engaging.

Concept 2:

A streamlined, AI-powered platform designed to assist users in transforming raw research materials into structured, polished outputs. This platform goes beyond simple search by helping researchers efficiently organize, synthesize, and present their findings in various formats, such as papers, presentations, or reports.

Concept 3:

The system helps creators generate mind maps, organize trending topics, and craft relevant posts and scripts with ease. By automatically curating top trends and topics, the platform presents a constantly updated list of categories on the homepage, allowing creators to effortlessly browse for inspiration.

Choosing a concept

After careful team discussion we decided to embark on concept 2.

Navigating users

Because we had three main users which were professionals, students, and content creators we needed to begin to examine their current pain points with modern GENAI platforms. We also wanted to understand their learning experiences and what works for them.

We believed this would give us important information of navigating our design according to what works best for our users.

Therefore, we conducted several semi-structured interviews with professionals, students and content creators.

Competitor Analysis

We also took it upon ourselves to examine direct and indirect competitors. We wanted to better understand the market we were breaking into and what industry standards were being used in other platforms.

Developing empathy

After our extensive research from interviews to competitor analysis we created an assumptive persona for each of our target users.

LET’s CREATE

After all of our previous steps we felt ready to commit to a detailed timeline of design.

Phase one

Objective: Create low-fidelity wireframes for each feature, focusing on user flow and usability.

Timeline: 2 weeks

Activities:

  • Develop wireframes for Prompt Formatting Mechanism, Dual Modes, Source Verification, and Idea Wall.

  • Conduct rapid user testing with a small group (5–10 users) to gather initial feedback.

  • Iterate on wireframes based on feedback, refining flow and structure.

PHASE TWO

Objective: Build interactive mid-fidelity prototypes to further test user interactions and navigation.

Timeline: 2 weeks

Activities:

  • Create clickable prototypes in Figma to allow users to simulate interactions in Map Mode, Card Mode, and other features.

  • Collect data on user confidence, usability, and preference for visual hierarchy in the Source Verification Mechanism.

  • Iterate designs based on findings, incorporating adjustments in layout, interaction paths, or information display.

Phase three

Objective: Develop high-fidelity designs and finalize the prototype for user testing.

Timeline: 2 weeks

Activities:

  • Apply branding and visual design elements to the prototypes, refining details and interactions.

  • Conduct usability tests with a larger user base to validate the design’s effectiveness in achieving learning outcomes.

  • Collect final feedback on user satisfaction, confidence, and the utility of features like the Idea Wall for creativity.

Our Design

We decided on creating and designing four main components that would make our GENAI product standout and resonate with our target audiences.

Dual Interaction Mode

Personalized learning paths can improve user engagement and effectiveness. While Card Mode aligns with traditional search-based learning, Map Mode facilitates a systematic, node-based exploration.


Prompt formatting

Studies highlight that knowledgeable users benefit most from time-saving features and that many users experience low confidence in their ability to use AI effectively. By offering guided prompt formatting, users can interact with AI more confidently and efficiently.


Source verification mechanism

Research shows users often lack confidence in AI-generated content, primarily due to uncertainty about its sources.

Inspiration Board

For users seeking to create content without a specific starting point, this feature offers a solution for quick, inspiration-based ideation.

The design

Our interpretation of all this information combined.

Design justification

Prompt Formatting:

Pre-set Prompts for Common Tasks: Provides ready-made templates for frequent tasks like summarizing, brainstorming, and categorizing. Users can customize these by filling in blanks or selecting keywords to tailor results to their needs.

Source Verification Method:

Enables users to specify the libraries (e.g., web search, academic databases, social media) for each query, offering flexibility based on their objectives. Also displays in-text citations of sources used to generate answers, ensuring transparency and reliability.

Dual Interaction Mode:

Node (Card) View: Expands a single node into a detailed, focused view. Users can dive deeper into one piece of information while maintaining a conversational flow with follow-up questions in a chat-like interface.

Automated Mind Mapping: Instantly generates a visual structure for a topic, allowing users to explore related ideas and dive deeper into areas of interest. This feature addresses the needs of structured learners by organizing information systematically.

Inspiration Board:

Category Navigation: Each idea card prominently displays its category, enabling users to click on it to explore related ideas. Upon selection, the idea hub dynamically refreshes to showcase all ideas within the same category, enhancing discoverability and contextual exploration.

  • Enhanced Visual Interest with Images: Each idea card includes a relevant image alongside the text to make it more visually engaging. These images provide context, spark creativity, and make the browsing experience more captivating and intuitive.

  • Infinite Flow Interaction: The idea hub adopts a social media-inspired layout, similar to Twitter, where idea cards are organized for effortless browsing. Users can continuously scroll to uncover a stream of fresh ideas, ensuring an uninterrupted and engaging discovery experience.

slide deck

Coming Soon….